Security audit

Web scraping skill using Chrome + WebMCP

Security checks for vulnerabilities and agentic risk

Overview

This scraping skill is not malicious, but it should be reviewed because it forces use of the user's live Chrome browser and may close tabs for broad search or scraping requests.

Install only if you are comfortable letting the skill control your live Chrome session. Use a dedicated browser profile with no sensitive logins, avoid invoking it for generic searches, and expect that its cleanup instructions may close tabs unless the agent constrains itself to tabs it opened.

A.I.G

Vulnerability Patterns

Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally

Findings (1)

T05 · Unauthorized Access and Privilege Escalation

Warning

Location: SKILL.md:112
Finding: Mandatory Host-Browser Execution Bypasses Sandbox Isolation
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The trigger phrases are broad enough to match routine requests such as 'search xxx' or '查一下xxx', causing the skill to activate for general browsing tasks without clear user intent for host-browser automation. In this skill's context, that is more dangerous because activation leads to controlling the host Chrome instance and potentially interacting with real user tabs and sessions.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The skill instructs use of target="host" and mandates cleanup that closes tabs, but it does not clearly warn users that it will control the host browser and may close existing tabs. This creates a real risk of destructive interference with the user's active browsing session, including loss of work, disruption of authenticated sessions, or unintended actions in a privileged browser context.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 78% confidence
Finding: Most trigger conditions and supported-site examples are tied to Chinese-language services such as Tonghuashun, East Money, Xueqiu, and Baidu, and the examples are primarily in Chinese. The document does not state that this locale focus is optional or region-specific, nor does it offer a language or locale choice to the user.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.